Enterprise business data interaction method and system based on AI intelligent customer service

Abstract

The present application relates to a method and system for interacting with enterprise business data based on AI intelligent customer service, which includes the steps of: when receiving a message to be identified sent by a client, sending the message to be identified to a preset emotion identification model; when the emotion identification model receives the message to be identified, matching the emotion identification associated with the message to be identified; based on the associated emotion identification and the message to be identified, matching the business process node associated with the current emotion identification and the message to be identified in the preset business process, and filtering out the reply element information associated with the business process node from the preset business information database; based on the pre-connected AI model, generating an associated reply message in combination with the reply element information and sending it to the client. The present application realizes the business proactive conversation capability of AI customer service, and improves the control of user emotions, so as to enhance the user experience in the AI ​​customer service process.

Description

Technical Field

[0001] The present application relates to the technical field of AI customer service emotion recognition, and in particular to a method and system for interacting with enterprise business data based on AI intelligent customer service. Background Art

[0002] At present, AI technology is widely used in the field of intelligent customer service. In the enterprise customer service sector, AI customer service robots are usually used to identify voice and text messages sent by users and perform semantic analysis, and generate reply texts through deep learning models to respond to customers.

[0003] However, the current AI customer service model only responds based on the content of user text messages or voice-converted text messages, that is, it remains at the text response stage and lacks the proactive business chat function like a salesperson. In addition, it is difficult to accurately control user emotions through text messages, and mechanical responses can easily reduce user experience. Therefore, this model needs to be improved. Summary of the invention

[0004] In order to realize the business proactive conversation capability of AI customer service and improve the control of user emotions, so as to enhance the user experience in the AI ​​customer service process, this application provides an enterprise business data interaction method and system based on AI intelligent customer service.

[0005] The above-mentioned invention objective of the present application is achieved through the following technical solutions:

[0006] A method for interacting with enterprise business data based on AI intelligent customer service, comprising the steps of:

[0007] When receiving a message to be identified sent by a client, sending the message to be identified to a preset emotion identification model;

[0008] When the emotion identification model receives the message to be identified, it matches the emotion identification associated with the message to be identified;

[0009] Based on the associated emotion identifier and the message to be identified, in a preset business process, a business process node associated with the current emotion identifier and the message to be identified is matched, and reply element information associated with the business process node is filtered out from a preset business information database;

[0010] Based on the pre-connected AI model, an associated reply message is generated in combination with the reply element information and sent to the client.

[0011] By adopting the above technical solution, during the service process of enterprise AI customer service, the user sends a message through the client. At this time, the preset emotion identification model obtains the message to be identified, wherein the message to be identified includes text, voice, etc. After semantic analysis of the message to be identified, the emotion identification representing the current user's emotion is matched, so that the enterprise AI customer service can obtain the dialogue context, thereby more accurately controlling the customer's emotion. Furthermore, according to the associated emotion identification and the message to be identified, the associated business process node is matched, and then the associated reply element information is matched. Finally, through the AI ​​model, the associated reply message is generated in combination with the reply element information. On the basis of emotion recognition, by combining the business process of the enterprise, on the basis of knowing the different emotions of the user, combined with its own business SOP, solutions are actively provided to the customer, and the business proactive conversation capability of the AI ​​customer service is realized. Moreover, the conversation is based on the customer emotion identification, which is closer to the real business personnel, thereby improving the control of the user's emotions, thereby improving the user experience in the AI ​​customer service process.

[0012] In a preferred example of the present application, when the emotion identification model receives the message to be identified, the step of matching the emotion identification associated with the message to be identified includes the following steps:

[0013] When the emotion identification model receives a message to be identified, the message type of the message to be identified is obtained, where the message type includes a text type;

[0014] When the message type is text type, semantic analysis is performed on the message to be identified and the associated emotional keywords are identified and extracted based on the preset extraction rules;

[0015] If an emotion keyword is extracted, then based on the emotion keyword and the semantic analysis result, a related emotion identifier is matched from a preset emotion identifier database.

[0016] By adopting the above technical solution, when the message type is text type, semantic analysis is performed on the identified message, and AI customer service can grasp the user's intentions and emotions more accurately. According to the preset emotional keyword extraction rules, such as identifying positive or negative emotional words in the text, the user's emotional state is judged. Through the matching of such emotional markers, AI customer service can more accurately control customer emotions.

[0017] In a preferred example of the present application: the message type also includes an audio type, and when the emotion identification model receives the message to be identified, the step of matching the emotion identification associated with the message to be identified also includes the steps of:

[0018] If the message type is an audio type, then obtaining audio attributes of the message to be identified, wherein the audio attributes include spectrum features, amplitude features, sound texture features, and pitch rhythm features;

[0019] Perform text conversion and semantic analysis on the message to be recognized, and extract related emotional keywords based on preset extraction rules;

[0020] Based on the semantic analysis results, emotion keywords and audio attributes, associated emotion identifiers are matched from a preset emotion identifier database.

[0021] By adopting the above technical solution, when the message type is audio type, AI customer service obtains the spectral characteristics, amplitude characteristics, sound texture characteristics and pitch rhythm characteristics in the message to be identified, so as to further and more comprehensively judge the user's emotional state, and match the associated emotion identifiers from the preset emotion identifier database based on the audio attributes, associated semantic analysis results, and associated emotion keywords. AI customer service can more accurately identify the customer's emotional state.

[0022] In a preferred example of the present application: when the emotion identification model receives the message to be identified, after the step of matching the emotion identification associated with the message to be identified, the following steps are performed:

[0023] When the emotion identifier is matched, the emotion identifier is bound to the client, and a preset binding time countdown is started. During the binding time, the client is temporarily bound to the matched emotion identifier;

[0024] If the set of identifiers corresponding to the emotion identifier of the client changes within the binding time, the emotion identifier most recently associated with the client is temporarily bound to the client.

[0025] By adopting the above technical solution, for example, when the matched emotion identifier is the angry identifier, the angry identifier is bound to the client, and the binding duration countdown starts. During the binding duration, the client temporarily binds to the matched angry identifier. When the identifier set corresponding to the angry identifier changes, if the latest associated emotion identifier is the happy identifier, the client temporarily binds to the latest associated happy identifier, and the binding countdown starts, and the above process continues. Through this process, AI customer service can more reasonably identify the customer's emotional state, and AI customer service is closer to real business personnel.

[0026] In a preferred example of the present application, based on the associated emotion identifier and the message to be identified, in a preset business process, a business process node associated with the current emotion identifier and the message to be identified is matched, and the reply element information associated with the business process node is screened out from a preset business information database, including the steps of:

[0027] Based on the associated message to be identified, in a preset business process, a set of associated business process nodes is screened out;

[0028] Based on the associated emotion identifier, matching a business process node corresponding to the emotion identifier in the associated business process node set;

[0029] Filter out the response element information associated with the business process node from the preset business information database.

[0030] By adopting the above technical solution, the business process node set associated with the message to be identified is screened out, and the associated business process nodes are further matched from the associated business process node set according to the emotion identifier, and finally the reply element information of the associated business process node is screened out, and the reply element information includes the reply logic. Through this process, AI customer service can understand the customer's problems and needs more accurately and reasonably, thereby providing more effective solutions; providing personalized replies and suggestions based on the customer's emotional state, increasing the personalization and relevance of the service.

[0031] In a preferred example of the present application: the message type also includes an image type, and the steps of matching a business process node associated with the current emotion identifier and the message to be identified in a preset business process based on the associated emotion identifier and the message to be identified, and filtering out the reply element information associated with the business process node from a preset business information database also include the steps of:

[0032] If the message type is an image type, image feature extraction is performed on the message to be identified to obtain a number of image features;

[0033] Based on the image features, identifying whether there is a product image matching the image features in a preset product image database;

[0034] If there is a matching product image, then the product information corresponding to the product image is output;

[0035] The output product information is combined with the context message to match the associated business process nodes, and the context message includes the identified emotion identifier.

[0036] By adopting the above technical solution, when the received message type is an image type, AI customer service will extract image features such as color, shape and texture. Using these features, AI customer service searches for matching images in the product database. Once a matching product image is found, the client will immediately display relevant product information such as name, description and price. At the same time, AI customer service will also analyze contextual messages, including emotional markers. In this way, the system can more effectively identify products, provide relevant information, and guide users to the appropriate business process nodes to enhance user experience.

[0037] In a preferred example of the present application, if there is a matching product image, the step of outputting the product information corresponding to the product image comprises the steps of:

[0038] If there is a matching product image, then further obtain the comparison results of several image features, the comparison results include abnormal information and normal information;

[0039] Product information is generated based on the product name, abnormal information and normal information pre-associated with the product image.

[0040] By adopting the above technical solution, when the input image matches the image in the product database, AI customer service analyzes the comparison results of the image features, which are divided into normal information and abnormal information. The system combines these comparison results with the preset information of the product image, including product name, specifications and expected features, to generate comprehensive product information, which helps to quickly and accurately identify and solve problems in customer service scenarios.

[0041] The second object of the invention is achieved by the following technical solutions:

[0042] An enterprise business data interaction system based on AI intelligent customer service, including:

[0043] The message receiving module is used to send the message to be identified to the preset emotion identification model when receiving the message to be identified sent by the client;

[0044] Matching module: used for matching the emotion identifier associated with the message to be identified when the emotion identifier model receives the message to be identified; based on the associated emotion identifier and the message to be identified, matching the business process node associated with the current emotion identifier and the message to be identified in the preset business process, and matching the associated reply element information from the preset business information database;

[0045] Reply message module: used to generate an associated reply message based on the pre-connected AI model and the reply element information and send it to the client.

[0046] By adopting the above technical solution, the message to be identified sent by the client is received, the emotion identifier associated with the message is matched, the associated business process node is further matched, and finally the associated reply element information is matched. Based on the reply element information, an associated reply message is generated and sent to the client. The AI ​​customer service can understand the customer's input more accurately and reply to the customer in a more humane way. At the same time, the emotion recognition capability of the AI ​​customer service is seamlessly integrated with the enterprise's workflow and business system, which can automate customer service and other key business processes, improve service efficiency and customer satisfaction.

[0047] The third objective of the present application is achieved through the following technical solutions:

[0048] A computer device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the enterprise business data interaction method based on AI intelligent customer service are implemented.

[0049] The fourth objective of the present application is achieved through the following technical solutions:

[0050] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a method for interacting with enterprise business data based on AI intelligent customer service.

[0051] In summary, the present application includes at least one of the following beneficial technical effects:

[0052] 1. On the basis of emotion recognition, by combining the business processes of the enterprise, knowing the different emotions of users, and combining its own business SOP, it actively provides solutions to customers, and realizes the business proactive conversation capability of AI customer service. The conversation is based on customer emotion identification and is closer to real business personnel, thus improving the control of user emotions and thus improving the user experience in the AI ​​customer service process.

[0053] 2. When the message type is text type, semantic analysis is performed on the identified message. AI customer service can grasp the user's intentions and emotions more accurately. According to the preset emotional keyword extraction rules, such as identifying positive or negative emotional words in the text, the user's emotional state is judged. Through the matching of such emotional markers, AI customer service can more accurately control customer emotions.

[0054] 3. When the message type is audio type, AI customer service obtains the spectrum characteristics, amplitude characteristics, sound texture characteristics and pitch rhythm characteristics in the message to further and more comprehensively judge the user's emotional state. Based on the audio attributes, related semantic analysis results, and related emotional keywords, the related emotional identifiers are matched from the preset emotional identifier database. AI customer service can more accurately identify the customer's emotional state.

[0055] 4. When the emotion identifier is matched, the emotion identifier is bound to the client, and the preset binding time countdown is started. During the binding time, the client is temporarily bound to the matched emotion identifier; if the identifier set corresponding to the client's emotion identifier changes during the binding time, the emotion identifier most recently associated with the client is temporarily bound to the client. Through this process, AI customer service can more reasonably identify the emotional state of customers, and AI customer service is closer to real business personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1This is a flowchart of an implementation of an enterprise business data interaction method based on AI intelligent customer service in the present application;

[0057] Figure 2 This is a flowchart for implementing step S20 in an embodiment of an enterprise business data interaction method based on AI intelligent customer service in the present application;

[0058] Figure 3 This is another implementation flow chart of step S20 in an embodiment of an enterprise business data interaction method based on AI intelligent customer service of the present application;

[0059] Figure 4 This is an implementation flow chart after step S20 in an embodiment of an enterprise business data interaction method based on AI intelligent customer service of the present application;

[0060] Figure 5 This is a flowchart for implementing step S30 in an embodiment of an enterprise business data interaction method based on AI intelligent customer service in the present application;

[0061] Figure 6 This is another implementation flow chart of step S30 in an embodiment of an enterprise business data interaction method based on AI intelligent customer service of the present application;

[0062] Figure 7 This is a flowchart for implementing step S306 in an embodiment of an enterprise business data interaction method based on AI intelligent customer service in the present application;

[0063] Figure 8 This is a principle block diagram of a computer device of the present application. DETAILED DESCRIPTION

[0064] The following is combined with Figure 1-8 This application is described in further detail.

[0065] In one embodiment, if Figure 1 As shown, the present application discloses a method for interacting with enterprise business data based on AI intelligent customer service, which specifically includes the following steps:

[0066] S10: When receiving a message to be identified sent by a client, sending the message to be identified to a preset emotion identification model;

[0067] In this embodiment, the client is a mobile terminal or PC provided for users to consult corporate business, ask questions, and initiate AI customer service dialogues. The message to be identified refers to a message sent by the user through the client, including but not limited to text messages, emoticons, voice messages, image messages, and video messages;

[0068] The preset emotion identification model refers to a model that is pre-set and trained, and can represent the emotions of the client user according to the matching of the message to be identified sent by the client.

[0069] Specifically, when a message to be identified is received from a client, the message to be identified is sent to a preset model of emotion identification for determining the emotion of the user of the client.

[0070] S20: When the emotion identification model receives the message to be identified, it matches the emotion identification associated with the message to be identified;

[0071] In this embodiment, the emotion identifier refers to a text identifier used to associate with the client and represent the current user emotion of the client. The emotion identifier includes a positive identifier and a negative identifier. The positive identifier represents that the current user emotion of the client is relatively positive, such as happiness, gratitude, calmness, enthusiasm, satisfaction, relief, etc. The negative identifier represents that the user emotion of the client is negative, such as sadness, irritability, anger, disappointment, fear, disgust, panic, etc.

[0072] Specifically, according to the message sent by the client, an emotion identifier representing the user emotion of the current client is matched in a preset model, and the associated emotion identifier may be one or more.

[0073] S30: Based on the associated emotion identifier and the message to be identified, in a preset business process, matching a business process node associated with the current emotion identifier and the message to be identified, and filtering out reply element information associated with the business process node from a preset business information database;

[0074] In this embodiment, the business process node refers to the business process link of the enterprise, that is, the business SOP of the enterprise. The business SOP of different enterprises can be customized;

[0075] The preset business information database refers to a pre-set database used to store reply element information. In the same business process node, different emotion identifiers correspond to different reply element information. For example, when a user reports a problem with a product through the client, and the emotion is determined to be a negative emotion of anger, the reply element information corresponding to the angry emotion identifier is selected, such as calming the user's emotions before answering the user's questions. For another example, if the user's emotion is determined to be satisfied, the reply element information corresponding to the emotion identifier representing satisfaction is filtered out, that is, when the user is satisfied, praise is given and other related business products are recommended.

[0076] The reply element information includes the reply keywords and reply logic. The reply logic includes the order of reply texts sent when controlling the AI ​​customer service to reply to the client.

[0077] Specifically, based on the current client's emotion identifier and the message to be identified sent by the client, the related enterprise business process links are matched in the preset enterprise's business process, and the reply element information associated with the current emotion identifier is matched from a pre-set database for storing reply element information to guide the reply text and reply order of the AI ​​customer service.

[0078] S40: Based on the pre-connected AI model, an associated reply message is generated in combination with the reply element information and sent to the client.

[0079] In this embodiment, the AI ​​model refers to a large language model pre-accessed by the AI ​​customer service, such as a chatgpt model.

[0080] Among them, when there is reply element information of reply logic, the reply keywords are sent to the AI ​​model in a logical order to realize the sequential output of the reply text.

[0081] Specifically, based on the large language model pre-accessed by the AI ​​customer service, an associated reply message is generated in combination with the reply element information and sent to the client as the reply text of the AI ​​customer service for replying to the client.

[0082] In one embodiment, referring to Figure 2 , step S20 comprises the steps of:

[0083] S201: When the emotion identification model receives a message to be identified, the message type of the message to be identified is obtained, where the message type includes a text type;

[0084] In this embodiment, the text types include words, emoticons, etc., and include the function of recognizing texts in different languages. The emotion identification model pre-translates the message to be identified into Chinese to match the emotion identification and the Chinese reply element information, where the reply element information includes the language information of the message to be identified, and then is identified and translated into the language corresponding to the message to be identified through the AI ​​model.

[0085] S202: When the message type is a text type, semantic analysis is performed on the message to be identified and associated emotional keywords are identified and extracted based on a preset extraction rule;

[0086] In this embodiment, the preset extraction rules are used to extract text and expression characters that can reflect the user's emotions, including special characters with a fixed arrangement order. The extracted emotion keywords should be judged in combination with semantic analysis.

[0087] Specifically, semantic analysis is performed on the text message and the associated emotional keywords are extracted based on a preset rule for extracting associated emotional keywords.

[0088] S203: If an emotion keyword is extracted, then based on the emotion keyword and the semantic analysis result, a related emotion identifier is matched from a preset emotion identifier database.

[0089] In this embodiment, semantic analysis can provide the context of the message to be identified, and further identify the intention of the emotional keywords in the context, so as to more accurately find the current emotional identifier of the client.

[0090] Specifically, when the associated emotion keywords are extracted, the context of the current message to be identified is determined based on the semantic analysis result, and the associated emotion keywords are combined with the context to match the associated emotion identifiers from a preset emotion identifier database.

[0091] When no emotional keywords are extracted, the emotional identifier associated with the client is determined through the semantic analysis result.

[0092] In one embodiment, referring to Figure 3 , step S20 further comprises the steps of:

[0093] S204: If the message type is an audio type, then obtaining audio attributes of the message to be identified, wherein the audio attributes include spectrum features, amplitude features, sound texture features, and pitch rhythm features;

[0094] In this embodiment, audio attributes refer to audio features. Through spectrum features, amplitude features, sound texture features, and pitch and rhythm features, the user's emotions can be objectively analyzed. For example, if the sound is continuously at a high frequency and has a high amplitude, it means that the user's voice is relatively high-pitched, and the user's speaking speed can be judged from the pitch and rhythm.

[0095] S205: performing text conversion and semantic analysis on the message to be recognized, and extracting associated emotional keywords based on preset extraction rules;

[0096] S206: Based on the semantic analysis results, emotion keywords and audio attributes, matching related emotion identifiers from a preset emotion identifier database.

[0097] In this embodiment, the context of the message to be identified by the current client is obtained based on the results of semantic analysis, and then the associated emotion identifiers are matched from the emotion identifier database through the associated emotion keywords and audio features. The emotion identifier of the client is jointly judged by combining the audio attributes, making the search for emotion identifiers more accurate.

[0098] In one embodiment, referring to Figure 4 , after step S20, perform the following steps:

[0099] S21: When the emotion identifier is matched, the emotion identifier is bound to the client, and a preset binding time countdown is started, during which the client is temporarily bound to the matched emotion identifier;

[0100] In this embodiment, when the emotion identifier is successfully matched with the client, the emotion identifier is associated with the client and a set time limit is activated. During this period of time, the client will temporarily retain this emotion identifier, that is, the AI ​​customer service can temporarily memorize the emotion identifier of the current client to prevent the unreasonable change of the emotion identifier in a short period of time, and continue to provide accurate reply text in the subsequent reply process to the client.

[0101] Furthermore, after the countdown ends, the associated emotion markers will be re-evaluated to maintain the relevance and accuracy of the service. This mechanism ensures that users receive a consistent and accurate service experience over a period of time.

[0102] S22: If the set of identifiers corresponding to the emotion identifier of the client changes within the binding time, the emotion identifier most recently associated with the client is temporarily bound to the client.

[0103] In this embodiment, the identification set refers to emotion identifications of the same type, including emotion identifications representing positive emotions as one identification set and emotion identifications representing negative emotions as one identification set.

[0104] If the client's emotion identifier changes during the binding time, and the change in the emotion identifier spans the identifier set, the associated emotion identifier of the user account will be automatically updated to ensure that the user's current emotional state is accurately reflected.

[0105] Furthermore, the new emotion logo will overwrite the previous one and reset the countdown of the binding time so as to continue to provide customer service that matches the user's current emotion. This dynamic update mechanism ensures that users can get the appropriate service experience throughout the interaction process.

[0106] In one embodiment, referring to Figure 5 , step S30 comprises the steps of:

[0107] S301: Based on the associated message to be identified, a set of associated business process nodes is screened out in a preset business process;

[0108] In this embodiment, the business process node set collects a variety of response element information filtering paths for users with different emotion identifications in the same business node. For example, the business node is the delivery time of product A, and the business process node set includes filtering paths for different response element information for the delivery time of product A when users have different emotions.

[0109] Specifically, based on the message to be identified, in a preset business process, a screening path for response element information corresponding to different emotion identifiers of the user in an associated business process node is screened out.

[0110] S302: Based on the associated emotion identifier, match the business process node corresponding to the emotion identifier in the associated business process node set;

[0111] Specifically, based on the associated emotion identifier, a screening path of specific response element information that matches the current emotion identifier of the client user is matched in the associated business process link set.

[0112] S303: Filter out the response element information associated with the business process node from the preset business information database.

[0113] Specifically, based on the screening path of the reply element information obtained by the emotion identification, the corresponding reply element information is extracted from the business information database.

[0114] For example, if the user is asking about product returns, the business process node set of the relevant return process, conditions, required documents and contact information is first found. At this time, if the user's emotion is marked as angry, the reply logic in the reply element information is to first provide soothing language text and quickly provide a solution.

[0115] In one embodiment, referring to Figure 6 , step S30 further comprises the steps of:

[0116] S304: If the message type is an image type, image feature extraction is performed on the message to be identified to obtain a number of image features;

[0117] S305: Based on the image features, identifying whether there is a product image matching the image features in a preset product image database;

[0118] S306: If there is a matching product image, output the product information corresponding to the product image;

[0119] S307: Matching the associated business process nodes based on the output product information and the context message, wherein the context message includes the identified emotion identifier.

[0120] In this embodiment, the product information includes the product name and the current status information of the product.

[0121] Product status information refers to the physical status information of the product in the image sent by the client, such as whether part A of product is damaged or part B of product is missing.

[0122] Specifically, by identifying the image sent by the client and performing feature comparison with the image in its own pre-stored product database, the information of the product sent by the client can be known. Among them, not only the name of the product can be known, but also the feature comparison can determine whether the product in the image is damaged, thereby improving the user experience of AI customer service.

[0123] Based on the product information of the output product image, the relevant business process links are matched in combination with the context message. The context message includes the emotion identifier that has been recognized by the system, so as to provide customers with more accurate services.

[0124] In one embodiment, referring to Figure 7 , step S306 includes the steps of:

[0125] S3061: If there is a matching product image, further obtain the comparison results of the respective image features, the comparison results including abnormal information and normal information;

[0126] In this embodiment, abnormal information includes damage, wear, flaws or other non-standard features in the image, and normal information includes the consistency and expected characteristics of the product;

[0127] Specifically, if a product image is matched in a pre-set database storing product images, several image features of the product image are further obtained and compared, and the comparison result includes abnormal information and normal information, wherein the abnormal information includes damage, wear, defects or other non-standard features in the image, and the normal information includes consistency and expected features of the product;

[0128] For example, in an after-sales service scenario, a user might upload a product image to report a potential defect or request technical support. Image analysis can identify the product and provide comprehensive information about its status, including any detected anomalies, thereby accelerating the problem-solving process and improving user satisfaction.

[0129] S3062: Generate product information based on the product name, abnormal information and normal information pre-associated with the product image.

[0130] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0131] In one embodiment, a business data interaction system based on AI intelligent customer service is provided, and the business data interaction system based on AI intelligent customer service corresponds to the business data interaction method based on AI intelligent customer service in the above embodiment. The business data interaction system based on AI intelligent customer service includes:

[0132] The message receiving module is used to send the message to be identified to the preset emotion identification model when receiving the message to be identified sent by the client;

[0133] Matching and screening module: used for matching the emotion identification associated with the message to be identified when the emotion identification model receives the message to be identified; based on the associated emotion identification and the message to be identified, matching the business process node associated with the current emotion identification and the message to be identified in the preset business process, and screening the reply element information associated with the business process node from the preset business information database;

[0134] Reply message module: used to generate an associated reply message based on the pre-connected AI model and the reply element information and send it to the client.

[0135] Optionally, also include

[0136] Message type module: used to obtain the message type of the message to be identified when receiving the message to be identified sent by the client. The message type includes text type, audio type, and picture type;

[0137] Text message analysis and extraction module: used for obtaining the message type of the message to be identified as a text type when receiving the message to be identified sent by the client; performing semantic analysis on the message to be identified; and extracting emotional keywords associated with the message to be identified based on preset extraction rules;

[0138] Text message matching module: used to match the associated emotion identifiers from a preset emotion identifier library based on the associated emotion keywords and the associated semantic analysis results.

[0139] Optionally, also include

[0140] Audio message module: when receiving a message to be identified from a client, obtain the message type of the message to be identified as an audio type, including spectrum features, amplitude features, sound texture features, and pitch rhythm features;

[0141] Audio message analysis and extraction module: performs text conversion and semantic analysis on the message to be identified; extracts emotional keywords associated with the message to be identified based on preset extraction rules;

[0142] Audio message matching module: used to match the associated emotion markers from a preset emotion marker library based on the associated semantic analysis results, associated emotion keywords and audio attributes.

[0143] Optionally, also include

[0144] Temporary binding module: when the emotion identifier is matched, the emotion identifier is bound to the client, and a preset binding time countdown is started. During the binding time, the client is temporarily bound to the matched emotion identifier;

[0145] Binding update module: If the set of identifiers corresponding to the emotion identifiers of the client changes within the binding period, the emotion identifier most recently associated with the client is temporarily bound to the client.

[0146] Optionally, also include

[0147] Module for filtering business process node collection: Based on the associated messages to be identified, the module matches the associated enterprise business process link collection in the pre-set business process;

[0148] Matching business process node module: Based on the associated emotion identifier, match the associated business process link in the associated business process link set;

[0149] Module for filtering reply element information: filtering reply element information associated with the business process node from a preset business information database.

[0150] Optionally, also include

[0151] Image message feature extraction module: used for obtaining the message type of the message to be identified as an image type when receiving the message to be identified sent by the client; performing image feature extraction on the message to be identified;

[0152] Identify and match product images module: Based on image features, identify whether there is a product image matching the image features in a preset product image database;

[0153] Output product information module: if there is a matching product image, then output the product information corresponding to the product image;

[0154] Matching business process node module: Matches the associated business process nodes based on the output product information combined with the context message, which includes the identified emotion identifier.

[0155] Optionally, also include

[0156] Image feature comparison module: If there is a matching product image, further obtain the comparison results of several image features, including abnormal information and normal information;

[0157] Generate product information module: Generate product information based on the product name, abnormal information and normal information pre-associated with the product image.

[0158] For the specific limitations of the enterprise business data interaction system based on AI intelligent customer service, please refer to the limitations of the enterprise business data interaction method based on AI intelligent customer service above, which will not be repeated here. Each module in the above-mentioned enterprise business data interaction system based on AI intelligent customer service can be implemented in whole or in part through software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0159] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 8 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store message types, emotion identifiers, product images, and product information. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, an enterprise business data interaction method based on AI intelligent customer service is implemented.

[0160] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements an enterprise business data interaction method based on AI intelligent customer service when executing the computer program.

[0161] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, a method for interacting with enterprise business data based on AI intelligent customer service is implemented.

[0162] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0163] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0164] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for interacting with enterprise business data based on AI intelligent customer service, characterized in that: Includes steps: When receiving a message to be identified sent by a client, sending the message to be identified to a preset emotion identification model; When the emotion identification model receives the message to be identified, it matches the emotion identification associated with the message to be identified; When the emotion identification model receives a message to be identified, the message type of the message to be identified is obtained, where the message type includes a text type; When the message type is text type, semantic analysis is performed on the message to be identified and the associated emotional keywords are identified and extracted based on the preset extraction rules; If an emotion keyword is extracted, then based on the emotion keyword and the semantic analysis result, a related emotion identifier is matched from a preset emotion identifier database; If the message type is an audio type, then obtaining audio attributes of the message to be identified, wherein the audio attributes include spectrum features, amplitude features, sound texture features, and pitch rhythm features; Perform text conversion and semantic analysis on the message to be recognized, and extract related emotional keywords based on preset extraction rules; Based on the semantic analysis results, emotional keywords and audio attributes, matching related emotional tags from a preset emotional tag database; Based on the associated emotion identifier and the message to be identified, in a preset business process, a business process node associated with the current emotion identifier and the message to be identified is matched, and reply element information associated with the business process node is filtered out from a preset business information database; Based on the associated message to be identified, in a preset business process, a set of associated business process nodes is screened out; Based on the associated emotion identifier, matching a business process node corresponding to the emotion identifier in the associated business process node set; Filtering out the reply element information associated with the business process node from a preset business information database; If the message type is an image type, image feature extraction is performed on the message to be identified to obtain a number of image features; Based on the image features, identifying whether there is a product image matching the image features in a preset product image database; If there is a matching product image, then the product information corresponding to the product image is output; Matching the associated business process nodes based on the output product information combined with the context message, where the context message includes the identified emotion identifier; Based on the pre-connected AI model, an associated reply message is generated in combination with the reply element information and sent to the client.

2. According to claim 1, a method for interacting with enterprise business data based on AI intelligent customer service is characterized by: When the emotion identification model receives the message to be identified, after the step of matching the emotion identification associated with the message to be identified, the following steps are performed: When the emotion identifier is matched, the emotion identifier is bound to the client, and a preset binding time countdown is started. During the binding time, the client is temporarily bound to the matched emotion identifier; If the set of identifiers corresponding to the emotion identifier of the client changes within the binding time, the emotion identifier most recently associated with the client is temporarily bound to the client.

3. According to claim 1, a method for interacting with enterprise business data based on AI intelligent customer service is characterized in that: The step of outputting product information corresponding to the product image if there is a matching product image comprises the steps of: If there is a matching product image, then further obtain the comparison results of several image features, the comparison results include abnormal information and normal information; Product information is generated based on the product name, abnormal information and normal information pre-associated with the product image.

4. An enterprise business data interaction system based on AI intelligent customer service, characterized in that: include: The message receiving module is used to send the message to be identified to the preset emotion identification model when receiving the message to be identified sent by the client; Matching and screening module: used for matching the emotion identification associated with the message to be identified when the emotion identification model receives the message to be identified; Based on the associated emotion identifier and the message to be identified, in a preset business process, a business process node associated with the current emotion identifier and the message to be identified is matched, and reply element information associated with the business process node is filtered out from a preset business information database; Text message module: used for obtaining the message type of the message to be identified when the emotion identification model receives the message to be identified, wherein the message type includes a text type; Text message analysis and extraction module: when the message type is text type, it performs semantic analysis on the message to be identified and identifies and extracts related emotional keywords based on preset extraction rules; A text message matching module: used to match the associated emotion identifiers from a preset emotion identifier database based on the emotion keywords and the semantic analysis results if emotion keywords are extracted; Audio message module: used to obtain the audio attributes of the message to be identified if the message type is audio type, wherein the audio attributes include spectrum characteristics, amplitude characteristics, sound texture characteristics and pitch rhythm characteristics; Audio message analysis and extraction module: used to convert text and perform semantic analysis on the message to be recognized, and extract related emotional keywords based on preset extraction rules; Audio message matching module: used to match the associated emotion identifier from a preset emotion identifier database based on the semantic analysis results, emotion keywords and audio attributes; A module for filtering a business process node set: used to filter out a set of associated business process nodes in a preset business process based on the associated message to be identified; A business process node matching module is used to match the business process node corresponding to the emotion identifier in the associated business process node set based on the associated emotion identifier; Filtering reply element information module: used to filter out reply element information associated with the business process node from a preset business information database; Image message feature extraction module: if the message type is an image type, then image feature extraction is performed on the message to be identified to obtain a number of image features; Identification and matching product image module: used to identify whether there is a product image matching the image features from a preset product image database based on the image features; Output product information module: used to output product information corresponding to the product image if there is a matching product image; Business process node matching module: used to match the associated business process nodes based on the output product information combined with the context message, where the context message includes the identified emotion identifier; Reply message module: used to generate an associated reply message based on the pre-connected AI model and the reply element information and send it to the client.

5. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the enterprise business data interaction method and system based on AI intelligent customer service are implemented as described in any one of claims 1 to 3.

6. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by the processor, the steps of the enterprise business data interaction method and system based on AI intelligent customer service are implemented as described in any one of claims 1 to 3.

Images